Quick, sharp and bite-sized, Radar distills the trending business-meets-culture stories from APAC to the Middle East to global markets into what you need to know. If it’s on our radar, it should be on yours.
OpenAI is ramping up efforts to move artificial intelligence beyond chatbots and into the real mechanics of business.
The company has reportedly launched Project Mercury according to several media reports, hiring over 100 former investment bankers from Goldman Sachs, JPMorgan Chase, and Morgan Stanley to help train its models on complex financial workflows. Reportedly paid around US$150 an hour, these bankers are teaching OpenAI’s systems how to perform real-world tasks such as IPO modelling, valuations, and M&A analysis, the technical groundwork that underpins global finance.
The goal is to embed professional reasoning into its models, creating AI that can replicate high-value human decision-making rather than just generate text. It’s an early sign that OpenAI is positioning itself to power domain-specific automation across industries, starting with finance.
At the same time, the company has introduced ChatGPT Atlas, a new web browser that merges browsing with conversation. The browser lets users highlight text, ask questions, and execute actions directly on any webpage. It includes Memory, which allows ChatGPT to retain context across sessions, and an Agent Mode that can plan trips, manage tasks, and perform multi-step operations. Currently available for macOS, with Windows, iOS, and Android to follow, Atlas effectively transforms ChatGPT from a destination into an operating environment, where search, work, and automation happen in one place.
BTB So What?
At first glance, this is simple automation, AI handling Excel models. In reality, it’s dismantling finance’s apprenticeship system. For decades, junior analysts learned by building models, catching errors, and doing grunt work under pressure. This wasn’t just skills training, it built instinct, judgment, and discipline. If AI assumes that foundational work, the industry loses its primary training ground. The next generation may advance faster but shallower: skilled at prompt engineering, not pattern recognition.
There’s also a cultural shift. Finance has always run on proprietary knowledge and guarded expertise. OpenAI’s integration systematises what was once human judgment, turning tacit skill into scalable infrastructure. Financial competence may soon depend less on who trained you and more on how well you engineer inputs. It starts with modeling. But the real story is this: the traditional path to expertise in finance is being rewritten. And once machines can replicate banker judgment, the question becomes whether bankers still need to think like machines, or if that cognitive model becomes obsolete entirely.
In parallel, Atlas represents a behavioural rewiring of how we interact with information. For two decades, browsers were passive windows; now they’re becoming co-workers. Instead of jumping between Google, Excel, and email, users can think, search, and act inside a single AI-mediated environment. It’s the end of task-switching and the beginning of continuous cognition, a workflow where the line between decision-making and execution disappears.
It starts with modeling. But the real story is this the traditional path to expertise in finance is being rewritten, while the fundamental structure of knowledge work itself is being collapsed into seamless execution. Once machines can replicate banker judgment and eliminate the friction of translation between thought and action, the question becomes whether the old model of professional development—and work itself—remains relevant at all.